I am a Software Engineer with a deep-rooted passion for building systems that matter — architecting scalable backends, shipping production-ready full-stack applications, and developing intelligent AI/ML solutions that solve real problems.
My engineering philosophy centers on clarity, scalability, and impact. I approach every system design decision with first-principles thinking, building for maintainability and performance at scale. With a strong foundation in computer science fundamentals and hands-on experience across the modern software stack, I bridge the gap between research-grade AI and production-grade engineering.
I specialize in Generative AI, Large Language Models, distributed systems, and cloud-native architectures — translating complex technical requirements into elegant, deployable solutions that drive measurable business outcomes.
Open To: Full-time SWE / MLE roles · AI/ML Engineering · Backend Systems · Full Stack Positions · Remote Opportunities · Open Source Collaboration
| Domain | Proficiency | Details |
|---|---|---|
| Large Language Models | ████████████ Expert | GPT-4, Claude, Gemini, Llama 3, fine-tuning, RLHF, prompt engineering |
| Retrieval-Augmented Generation | ████████████ Expert | LangChain, LlamaIndex, vector stores, hybrid search, reranking |
| Computer Vision | █████████░░░ Advanced | CNNs, YOLO, image segmentation, OpenCV, feature extraction |
| Natural Language Processing | █████████░░░ Advanced | Transformers, BERT, semantic search, text classification, NER |
| ML Ops & Pipelines | ████████░░░░ Proficient | MLflow, model versioning, A/B testing, inference optimization |
| Generative AI | ████████████ Expert | Diffusion models, multi-modal AI, AI agents, tool-use patterns |
| Deep Learning Frameworks | █████████░░░ Advanced | PyTorch, TensorFlow, Keras, Hugging Face Transformers |
| Vector Databases | ████████░░░░ Proficient | Pinecone, Weaviate, ChromaDB, FAISS, pgvector |
⬡ IntelliQuery — AI-Powered Enterprise Search Platform
A production-grade Retrieval-Augmented Generation (RAG) platform designed for enterprise knowledge bases. Enables natural language querying over internal documents, PDFs, and structured data with sub-second latency and hallucination mitigation pipelines.
| Attribute | Details |
|---|---|
| Stack | Python · FastAPI · LangChain · OpenAI GPT-4 · Pinecone · PostgreSQL · React · Docker |
| Scale | Handles 50,000+ document ingestions · 10K+ daily queries · multi-tenant architecture |
| Performance | P99 query latency < 800ms · 94% retrieval precision · chunk-level citation accuracy |
| Security | JWT auth · RBAC · encrypted vector store · rate limiting · audit logging |
| Impact | Reduced enterprise search time by 68% · adopted by 3 internal teams in pilot |
| Repository |  |
IntelliQuery solves the fundamental problem of knowledge silos in organizations. Built with a hybrid retrieval architecture combining dense vector search with BM25 sparse retrieval, it achieves state-of-the-art recall across heterogeneous corpora. The system includes automatic document chunking strategies, metadata-aware reranking, and a confidence-based response filtering pipeline that ensures answer reliability in high-stakes environments.
⬡ NexusAPI — Distributed Microservices Gateway
A cloud-native API gateway and microservices orchestration layer built for high-throughput production workloads. Designed with event-driven architecture, circuit breakers, and zero-downtime deployment capabilities.
| Attribute | Details |
|---|---|
| Stack | Node.js · Go · Kafka · Redis · PostgreSQL · Kubernetes · AWS EKS · Prometheus · Grafana |
| Scale | 1M+ API requests/day · 12 microservices · 99.97% uptime SLA |
| Performance | Avg gateway latency 45ms · horizontal pod autoscaling · L7 load balancing |
| Security | mTLS inter-service communication · OAuth2 · API key rotation · DDoS protection |
| Impact | Reduced infrastructure cost by 34% vs monolith · deployment frequency 10× |
| Repository | |
NexusAPI was built to address the operational complexity of large-scale service meshes. The gateway handles dynamic routing, request transformation, and observability instrumentation transparently, so downstream services remain clean and focused. The event-driven Kafka backbone enables fully asynchronous workflows with guaranteed delivery, while the Go-based proxy core ensures minimal CPU overhead even at peak traffic. Circuit breaker patterns prevent cascading failures across dependent services.
⬡ VisualCortex — Real-Time Computer Vision Pipeline
An end-to-end computer vision inference platform delivering real-time object detection, scene understanding, and visual analytics through a scalable streaming pipeline and REST/WebSocket API layer.
| Attribute | Details |
|---|---|
| Stack | Python · PyTorch · YOLO v8 · FastAPI · OpenCV · WebSocket · Redis Streams · AWS Lambda |
| Scale | 60 FPS real-time inference · batch processing 100K+ images/hour · serverless scaling |
| Performance | mAP@0.5 of 91.3% on custom dataset · model quantization (INT8) for 3× speedup |
| Security | Presigned S3 URLs · input validation · inference sandboxing · CORS hardening |
| Impact | Powers automated QA pipeline · reduced manual inspection overhead by 80% |
| Repository | |
VisualCortex abstracts the complexity of deploying CV models in production environments. The inference engine supports dynamic model swapping without downtime, enabling A/B testing of model versions in production. A WebSocket-based streaming layer allows real-time annotation overlays for live video feeds, while a Redis Streams backbone decouples ingestion from processing, providing backpressure-safe throughput even during traffic spikes.
⬡ Sentient — Full Stack AI SaaS Platform
A multi-tenant SaaS application offering AI-assisted workflows for productivity, content generation, and decision intelligence. Built with modern full-stack architecture and designed for enterprise-grade reliability.
| Attribute | Details |
|---|---|
| Stack | Next.js · TypeScript · FastAPI · PostgreSQL · Stripe · OpenAI API · Tailwind CSS · Vercel |
| Scale | 500+ active users · 3 subscription tiers · multi-org workspace support |
| Performance | Core Web Vitals green · streaming SSR · edge function response < 60ms |
| Security | Row-level security (Supabase) · encrypted secrets · OWASP compliance · CSP headers |
| Impact | $0 to live product in 6 weeks · 4.8/5 user satisfaction score in beta |
| Repository | |
Sentient was designed around the principle that AI features should feel native, not bolted on. Every workflow in the platform is augmented with contextual AI assistance — not as a gimmick, but as a productivity multiplier. The billing integration with Stripe handles metered usage accurately, and a feature-flag system allows safe rollout of new AI capabilities to percentage cohorts. The frontend is fully server-rendered with React Server Components, achieving excellent SEO and time-to-interactive metrics.
Software Engineering Intern · Neurovia Nexus Pvt Ltd
Jan 2024 – Present
B.Sc. Computer Science · University of the People
Jan 2024 – Present
Contributed to production-grade software systems and AI research initiatives within a fast-paced engineering environment, collaborating across cross-functional teams to deliver high-impact features from design to deployment.
- Designed and shipped RESTful microservices handling 100K+ daily requests with <50ms P95 latency using FastAPI and PostgreSQL
- Built and deployed ML inference pipelines for NLP classification tasks, achieving 89% F1 score on production datasets
- Implemented CI/CD workflows with GitHub Actions and Docker, reducing deployment cycle time from hours to under 12 minutes
- Collaborated on architecture decisions for a distributed cache layer using Redis, improving read throughput by 3×
- Authored internal engineering documentation and participated in bi-weekly code reviews across a 10-person team
| 🏆 Recognition | Details |
|---|---|
| Smart India Hackathon Finalist | National-level government hackathon · top 50 teams from 5,000+ submissions |
| Open Source Contributor | Merged PRs in public repositories · active GitHub contributor |
| AI/ML Research Project | Published internal report on RAG optimization for low-resource languages |
| Academic Excellence | Consistent top-quartile academic performance in Computer Science program |
| Technical Lead | Led a team of 6 engineers in college-level systems design competition |
| Dean's List Recognition | Recognized for outstanding academic and project contribution record |
madhurjya_bordoloi:
current_focus:
learning:
- Advanced LLM fine-tuning and alignment techniques (DPO, ORPO)
- Distributed systems design at scale (consensus algorithms, CRDTs)
- Rust for systems programming and WebAssembly targets
- Kubernetes operator development and custom resource definitions
building:
- Open source RAG evaluation framework for low-resource language benchmarks
- Personal AI assistant with long-term memory and tool-use capabilities
- Microservices orchestration template with observability baked in
- Full stack SaaS boilerplate with AI feature integrations
exploring:
- Multi-agent AI architectures (AutoGen, CrewAI, LangGraph)
- Inference optimization — quantization, speculative decoding, vLLM
- Edge AI deployment on resource-constrained environments
- Graph neural networks for relational data reasoning
open_to:
- Full-time SWE / MLE positions (India or Remote)
- AI/ML research collaborations
- Open source project contributions
- Technical mentorship and knowledge exchange
- Hackathons and engineering challenges
"The best engineers are not those who write the most code — but those who solve the right problems with the least complexity."